Navidgator - Similarity Based Browsing for Image & Video Databases
|
|
- Paul Brown
- 5 years ago
- Views:
Transcription
1 Navidgator - Similarity Based Browsing for Image & Video Databases Damian Borth, Christian Schulze, Adrian Ulges, and Thomas M. Breuel University of Kaiserslautern, German Research Center for Artificial Intelligence (DFKI), Kaiserslautern, Germany {d_borth,a_ulges,tmb}@informatik.uni-kl.de {christian.schulze}@dfki.de Abstract. A main problem with the handling of multimedia databases is the navigation through and the search within the content of a database. The problem arises from the difference between the possible textual description (annotation) of the database content and its visual appearance. Overcoming the so called - semantic gap - has been in the focus of research for some time. This paper presents a new system for similaritybased browsing of multimedia databases. The system aims at decreasing the semantic gap by using a tree structure, built up on balanced hierarchical clustering. Using this approach, operators are provided with an intuitive and easy-to-use browsing tool. An important objective of this paper is not only on the description of the database organization and retrieval structure, but also how the illustrated techniques might be integrated into a single system. Our main contribution is the direct use of a balanced tree structure for navigating through the database of keyframes, paired with an easy-to-use interface, offering a coarse to fine similarity-based view of the grouped database content. Key words: hierarchical clustering, image databases, video databases, browsing, multimedia retrieval 1 Introduction Nowadays, content-based image and video retrieval (CBIR/CBVR) are getting more and more into focus with the rapidly growing amount of image and video information being stored and published. Online portals providing image and video content like flickr.com, youtube.com, revver.com, etc. offer data in large amounts which creates the need for an efficient way of searching through the content. But the need for an efficient search and browsing method is not bound to those online portals. Archives of TV stations storing increasingly more digital content also need appropriate tools to find the material they are looking for. Additionally, with the already widely spread availability of digital acquisition
2 devices (still image and video cameras) it is easy for everybody to acquire large amounts of digital data in short amounts of time. Currently the search for specific content in such collections is mostly done through a query-by-text approach, exploiting manual annotation of the stored data. This approach suffers from a few drawbacks which arise from the nature of this method. First: manual annotation is a very time consuming process which second: might lead to a rather subjective result, depending on the person doing the annotation. Third: the result of the query depends highly on the quality of the annotations. Visual content that has not been transcribed into the metadata can therefor not be retrieved afterwards. Fourth: The result of a query can be manipulated by the type and number of tags associated with the visual data. This might mostly apply to online portals where currently the owners of uploaded content provide the meta-data. With such tag manipulation it is possible to assure that a specific content appears in most of the query results, which then in turn reduces the quality of the search result. To prevent such drawbacks it is necessary to use different approaches for search in large multimedia databases. One of these approaches is the description of the database content in a specific feature space. Unfortunately, a major problem of this approach is the formulation of a query. This is bypassed by using the query-by-example approach [1], where a sample image or video is representing a query forming a visual concept for search. Our approach is to build up a similarity based structure of the multimedia database (here we focus on video as content) to overcome the need for an appropriate search example. Doing this, the user is enabled to browse through the database by picking an entity as starting point which represents the query most. During browsing the user is able to zoom in and out of the database content with variable step size representing the similarity of appearance. Entities showing up during the navigation, giving a better representation of the users query, provide the opportunity to narrow down the selection of possible matches. 1.1 Related Work With the development of multimedia retrieval/browsing systems quickly the problem of formulating a proper query arose. This lead to the definition of the so called semantic gap - a lack of coincidence between the information that one can extract from the visual data and the interpretation that the same data have for a user in a given situation. [2]. Different approaches have been examined to bridge this semantic gap.[3 6] and lead to multiple approaches towards user interface design i.e. the RotorBrowser [7] or VideoSOM [8]. When we talk about similarity based browsing, we focus on the visual content of the video. This means we want to allow the user to find temporally independent shots of video with similar content, which is different to clustertemporal browsing where causal relations in storytelling are used for browsing [9]. We purely utilize content based similarity between different videos in the video database for browsing the database. A similar approach can be found in [10, 11], where the concept of a similarity pyramid is introduced for browsing
3 large image databases. The idea of similarity pyramids was also applied to video databases in a system called ViBE [12]. Our main contribution is the direct use of a balanced tree structure for navigating through the database of keyframes, paired with an easy to use interface, offering a coarse to fine view on the grouped database content based on similarity. 2 Balanced Hierarchical Clustering To find structures of strong visual similarity we use unsupervised learning methods like clustering. In particular we are not only interested in the pure cluster partitioning, even more we want to capture the relationship between different clustering levels i.e. the clustering structure in the video database including their cluster and their subcluster partitions. To achieve this we use a standard agglomerative hierarchical clustering [13], which runs through a series of partitions starting from singelton clusters containing a single image and terminating with a final cluster, containing all images of the database. This structure, usually represented as a dendogram, is postprocessed into a binary tree and used by our system for continuous browsing between different coarseness levels of similarity. Fig. 1. A feature space representation of keyframes (a) leads to a hierarchical clustering (b). This clustering can be viewed as a dendogram with similarity measurement at y- axis (c) and will be postprocessed to a binary tree for our GUI (d) 2.1 Feature Extraction Let the videos in the database be denoted as X 1,..., X n. Every video is represented by a set of keyframes {x i1,..., x im }, resulting in a total keyframe set {x 1,..., x k } for the entire video database. For every keyframe x i {x 1,..., x k } a feature vector z i R D is extracted. A common way to extract keyframes for each video is to segment the data into shots and analyze the shots individually for representative keyframes. We use a divide-and-conquer approach that delivers multiple keyframes per shot in respect to its visual complexity. To achieve this, we compute MPEG-7 color layout descriptors [14] for each frame of the shot and fit a Gaussian mixture model to the feature set using k-means [15] in combination with the Bayesian
4 Information Criterion (BIC) for determining the number of clusters. [16]. The entire procedure is illustrated in more detail in [17]. The computation of the feature vector z i for every keyframe x i {x 1...x k } is based on our Baseline System definition [18]. There we are using Color and Texture features with a equally weighted early fusion (i.e concatenated). In particular, we use color histograms which are quantized to bins and also Tamura texture features [19] to form the feature vector. 2.2 Agglomerative Hierarchical Clustering The first step in using conventional agglomerative hierarchical clustering algorithm is the computation of a distance matrix D = [d(z i, z j )] where i, j = 1...k, with k the number of keyframes. The distance matrix D represents similarities between all pairs of keyframes and is used for successive fusion of clusters [13]. As distance function d(z i, z j ) serves the Euclidian distance. The agglomerative hierarchical clustering creates a cluster c i = {x i } for each keyframe x i {x 1...x k }, resulting in C 0 = {c 1...c k } disjoint singelton clusters, where C 0 C. In this first step the distance matrix D is equal to the distances between the feature vectors of the keyframes. The cluster c i c j with the smallest distance d(z i, z j ) are fused together to form a new cluster c k = {c i, c j }. After creating a new cluster, the distance matrix D has to be updated to represent the distance between the new cluster c k to each other cluster in C 0 \{c k }. This recalculation of distances is usually done with one of the known linkage methods [11]. Considering the used linkage method the entire clustering structure will be more or less dilating i.e. individual elements not yet grouped are more likely to form new groups instead of being fused to existing groups. According to [11], the complete linkage method tends to be dilating and therefore resulting in more balanced dendograms compared to the single linkage method, which chains clusters together and therefore results in deep unbalanced dendograms. However, our goal is to use the resulting clustering structure for continuous similarity browsing of video databases. In order to achieve this, two points are important: First, the clustering must produce clusters with visual similar content and Second, the dendrogram produced by the clustering has to be as balanced as possible.according to our experience, average linkage produces the visually most similar clusters. Unfortunately, the clustering structure is not as balanced as desired for usable browsing. Therefore a modification of the average linkage method was needed to achieve the desired properties. Viewing the resulting dendogram of the clustering as a tree structure, this structure will hierarchically organize keyframes into groups of visual similar content, thereby retaining the relationship between different coarseness levels of similarity and tree depth. Let S denote the set of all tree nodes. Each node of the tree s S is associated with a set of keyframes c s C c s / C 0 representing the cluster of keyframes. The number of elements in the cluster c s is denoted by c s. The children of a node s S denoted by ch(s) S will partition the
5 keyframes of the parent node so that c s = r ch(s) The leaf nodes of the tree correspond to the extracted keyframes and are indexed by the set S 0. Each leaf node contains a single keyframe, so for all s i S 0 we have c i = {x i } with c i = 1 implying that S 0 = C 0. This notation is derived from the notations of [10 12]. Furthermore we define D = [d(c i, c j )] as the updated distance matrix of distances between the pairwise different clusters c i and c j. The linkage method for calculating distances between clusters with c i > 1 c j > 1 is in our case the average linkage method enhanced by a weighted penalty, which depends on the amount of elements in both clusters d(c i, c j ) = 1 c i c j z i c i c r z j c j d(z i, z j ) + α ( c i + c j ) We are naming this method: balanced linkage due to its tendency to form balanced trees with clusters of consistent visual properties. The weight can be set to 0 α 1, which either results in having no effect to the average linkage or totally forcing the algorithm to produce balanced trees without any visual similarity. The chosen α value was empirically evaluated and set to α = Binary Tree Construction The usually chosen representation of hierarchical clustering is a dendogram, which illustrates the fusions made at each successive stage of analysis. In a dendogram, the elements being clustered are placed usually at the bottom of the diagram and show fusions of clusters through connecting lines. Another representation for hierarchical clustering is using sets showing the elements being clustered in their feature space. Such sets represent one particular step in the clustering process and may contain subsets illustrating previous clustering steps. Because we want to use the clustering structure for navigation, we postprocess the dendogram into a binary tree. The binary tree structure enables us to efficiently follow a keyframe from the root of the tree, which contains the entire database, to the leaf of the tree, which only contains the selected keyframe. With every zoom in step, the system is presenting a more similar subtree considering the selected keyframe and leaving out the frames, which are less similar to the selected keyframe. The zoom out step, lets the system present a tree containing more dissimilar keyframes, which might be useful for refinement of the query. For the notation of the constructed binary tree, we refer to section 2.2. Let T = S be the binary tree, then the nodes n i T are the fusion points where clusters c i,c j are being fused together to form a new cluster c k = {c i, c j }. An interesting outcome of the binary tree postprocessing is the creation of so called content stripes. These structures represent clusters within the binary
6 tree, in such a way that the keyframes of subclusters are ordered in stripes according to their similarity and therefore providing a more intuitive way of visualizing clusters at a particular level (Fig. 2). Content stripes replace the need to additionally compute spatial arrangemets for keyframes within a cluster like done with similarity pyramides [10, 11]. Therefore our method is not only able to construct a hierarchical database structure but also to build up a similarity based order within a cluster in one single step instead of separating these tasks. Fig. 2. A binary tree representation of a sample clustering, transformed to a content stripe displaying similarity clusters in an one-dimensional order 3 Graphical User Interface Retrieval systems based on keyframes and best-match similarity tend to present a localized view of the database to the user, rather then providing an overview of the entire database. For uses who do not clearly know what exactly they are searching for, it would be more efficient to let them browse through the database and allow them to dynamically redefine their search query. In this section we would like to present the Navidgator graphical user interface 1, which allows a user to easily and efficiently browse a video database in respect to his selected query. In our system a browsing process is initialized with starting at the root of the hierarchical clustered binary tree. First, the user has to select his first keyframe out of a randomly sampled set from the entire database to formulate his query [1]. This keyframe will represent the visual concept, which will guide the user while browsing. The user is also able to dynamically refine his visual concept in every point during browsing by selecting another keyframe. Browsing itself is performed by the given zooming tools. The user can either zoom-in or zoom-out in the database. A zoom-in action will narrow down the available keyframe according to his visual concept and a zoom-out action will display a coarser level of the database to the user. For better usability the 1 ( keyframes database)
7 interface provides a multi-level-zoom-in action and a multi-level-zoom-out action. Furthermore the user can perform a max-zoom-in action, which brings him straight to the most similar keyframes in the database or a max-zoom-out action, which brings him back to the root of the binary tree i.e. the top of the database. The depth of the database and the users current position are visualized by a vertical bar next to the zooming tools enabling an intuitive orientation. Additionally the user can utilize a click history to jump back to particular points of his browsing process. The Navidgator browsing interface is displayed in Fig. 3. Fig. 3. The Navidgator browsing interface. The selected visual concept is displayed at the center. The lower area displays the cluster preview box, where the visual concept might be refined. Next to the selected keyframe the navigation tools are arranged 4 Conclusion In this paper we have presented a new system for similarity-based browsing of multimedia databases. By using a balanced tree based on hierarchical clustering of the database content it is possible to supply users with an intuitive and easyto-use browsing tool. We were able to improve search results by providing a set of navigation tools which support the decision tree like structure of the clustering. Because of our concept of an offline clustering and online retrieval we are able to efficiently perform a search on the entire database. The system offers coarse and detailed views on the database content with the opportunity to change the focus of search at any time. This enables the user to start navigation with a fuzzy visual concept and improve relevance incrementally while browsing. Our future work will focus on advanced tree building and on dealing with growing databases, which will basically cover merging of new data into the binary tree. Additionally, to bridge the semantic gap and include high-level semantics we want to enhance the browser with an automatic tagging system like in [18]
8 References 1. Flickner, M., Sawhney, H., Niblack, W., Ashley, J., Huang, Q., Dom, B., Gorkani, M., Hafner, J., Lee, D., Petkovic, D., Steele, D., Yanker, P.: Query by image and video content: the qbic system. Computer 28,Issue 9 (Sept. 1995) Smeulders, A., Worring, M., Santini, S., Gupta, A., Jain, R.: Contentbased image retrieval at the end ofthe early years. IEEE transactions on pattern analysis and machine intelligence 22(12) (2000) Broecker, L., Bogen, M., Cremers, A.B.: Bridging the semantic gap in contentbased image retrieval systems. In: Internet Multimedia Management Systems II. Volume 4519 of Presented at the Society of Photo-Optical Instrumentation Engineers (SPIE) Conference. (2001) Zhao, R., Grosky, W.: Narrowing the semantic gap-improved text-based web document retrieval using visual features. Multimedia, IEEE Transactions on 4(2) (2002) Zhao, R., Grosky, W.: Bridging the Semantic Gap in Image Retrieval. Distributed Multimedia Databases: Techniques and Applications (2003) 6. Dorai, C., Venkatesh, S.: Bridging the semantic gap with computational media aesthetics. Multimedia, IEEE 10(2) (2003) de Rooij, O., Snoek, C.G.M., Worring, M.: Mediamill: semantic video search using the rotorbrowser. [7] Barecke, T., Kijak, E., Nurnberger, A., Detyniecki, M.: Videosom: A som-based interface for video browsing. IMAGE AND VIDEO RETRIEVAL, PROCEEDINGS 4071 (2006) Rautiainen, M., Ojala, T., Seppanen, T.: Cluster-temporal browsing of large news video databases. IEEE Int. Conference on Multimedia and Expo 2 (2004) Chen, J., Bouman, C., Dalton, J.: Similarity pyramids for browsing and organization of large image databases. SPIE Human Vision and Electronic Imaging III 3299 (1998) 11. Chen, J.Y., Bouman, C., Dalton, J.: Hierarchical browsing and search of large image databases. Image Processing, IEEE Transactions on 9, Issue 3 (March 2000) Taskiran, C., Chen, J., Albiol, A., Torres, L., Bouman, C., Delp, E.: Vibe: A compressed video database structured for active browsing and search. IEEE Transactions on Multimedia 6(1) (FEB 2004) Johnson, S.C.: Hierarchical clustering schemes. PSYCHOMETRIKA 32(3) (1967) Manjunath, B., Ohm, J., Vasudevan, V., Yamada, A.: Color and texture descriptors. IEEE Trans. on Circuits Syst. for Video Techn. 11(6) (2001) 15. McQueen, J.: Some methods for classification and analysis of multivariate observations. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability (1967) Schwarz, G.: Estimating the dimension of a model. Annals of Statistics 6(2) (1978) Borth, D., Ulges, A., Schulze, C., Breuel, T.M.: Keyframe extraction for video taggging and summarization. In: Informatiktage (2008) Ulges, A., Schulze, C., Keysers, D., Breuel, T.M.: Content-based video tagging for online video portals. In: MUSCLE/Image-CLEF Workshop. (2007) 19. Tamura, H., Mori, S., Yamawaki, T.: Textual features corresponding to visual perception. IEEE Transactions on Systems, Man, and Cybernetics SMC-8(6) (1978)
Navidgator. Similarity Based Browsing for Image & Video Databases. Damian Borth, Christian Schulze, Adrian Ulges, Thomas M. Breuel
Navidgator Similarity Based Browsing for Image & Video Databases Damian Borth, Christian Schulze, Adrian Ulges, Thomas M. Breuel Image Understanding and Pattern Recognition DFKI & TU Kaiserslautern 25.Sep.2008
More informationVideo Key-Frame Extraction using Entropy value as Global and Local Feature
Video Key-Frame Extraction using Entropy value as Global and Local Feature Siddu. P Algur #1, Vivek. R *2 # Department of Information Science Engineering, B.V. Bhoomraddi College of Engineering and Technology
More informationCOMPARISON OF SOME CONTENT-BASED IMAGE RETRIEVAL SYSTEMS WITH ROCK TEXTURE IMAGES
COMPARISON OF SOME CONTENT-BASED IMAGE RETRIEVAL SYSTEMS WITH ROCK TEXTURE IMAGES Leena Lepistö 1, Iivari Kunttu 1, Jorma Autio 2, and Ari Visa 1 1 Tampere University of Technology, Institute of Signal
More informationUnsupervised Learning and Clustering
Unsupervised Learning and Clustering Selim Aksoy Department of Computer Engineering Bilkent University saksoy@cs.bilkent.edu.tr CS 551, Spring 2009 CS 551, Spring 2009 c 2009, Selim Aksoy (Bilkent University)
More informationA System that Learns to Tag Videos by Watching Youtube
A System that Learns to Tag Videos by Watching Youtube Adrian Ulges 1,2, Christian Schulze 2, Daniel Keysers 2, Thomas M. Breuel 1,2 1 Department of Computer Science, Technical University of Kaiserslautern
More informationUsing Natural Clusters Information to Build Fuzzy Indexing Structure
Using Natural Clusters Information to Build Fuzzy Indexing Structure H.Y. Yue, I. King and K.S. Leung Department of Computer Science and Engineering The Chinese University of Hong Kong Shatin, New Territories,
More information10601 Machine Learning. Hierarchical clustering. Reading: Bishop: 9-9.2
161 Machine Learning Hierarchical clustering Reading: Bishop: 9-9.2 Second half: Overview Clustering - Hierarchical, semi-supervised learning Graphical models - Bayesian networks, HMMs, Reasoning under
More informationUnsupervised Learning and Clustering
Unsupervised Learning and Clustering Selim Aksoy Department of Computer Engineering Bilkent University saksoy@cs.bilkent.edu.tr CS 551, Spring 2008 CS 551, Spring 2008 c 2008, Selim Aksoy (Bilkent University)
More informationImage Querying. Ilaria Bartolini DEIS - University of Bologna, Italy
Image Querying Ilaria Bartolini DEIS - University of Bologna, Italy i.bartolini@unibo.it http://www-db.deis.unibo.it/~ibartolini SYNONYMS Image query processing DEFINITION Image querying refers to the
More informationSketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix
Sketch Based Image Retrieval Approach Using Gray Level Co-Occurrence Matrix K... Nagarjuna Reddy P. Prasanna Kumari JNT University, JNT University, LIET, Himayatsagar, Hyderabad-8, LIET, Himayatsagar,
More informationHolistic Correlation of Color Models, Color Features and Distance Metrics on Content-Based Image Retrieval
Holistic Correlation of Color Models, Color Features and Distance Metrics on Content-Based Image Retrieval Swapnil Saurav 1, Prajakta Belsare 2, Siddhartha Sarkar 3 1Researcher, Abhidheya Labs and Knowledge
More informationA Survey on Content Based Image Retrieval
A Survey on Content Based Image Retrieval Aniket Mirji 1, Danish Sudan 2, Rushabh Kagwade 3, Savita Lohiya 4 U.G. Students of Department of Information Technology, SIES GST, Mumbai, Maharashtra, India
More informationImage retrieval based on bag of images
University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 2009 Image retrieval based on bag of images Jun Zhang University of Wollongong
More informationCS 2750 Machine Learning. Lecture 19. Clustering. CS 2750 Machine Learning. Clustering. Groups together similar instances in the data sample
Lecture 9 Clustering Milos Hauskrecht milos@cs.pitt.edu 539 Sennott Square Clustering Groups together similar instances in the data sample Basic clustering problem: distribute data into k different groups
More informationIdentifying Relevant Frames in Weakly Labeled Videos for Training Concept Detectors
Identifying Relevant Frames in Weakly Labeled Videos for Training Concept Detectors Adrian Ulges Christian Schulze Department of Computer German Research Center for Science, Technical University Artificial
More informationScalable Hierarchical Summarization of News Using Fidelity in MPEG-7 Description Scheme
Scalable Hierarchical Summarization of News Using Fidelity in MPEG-7 Description Scheme Jung-Rim Kim, Seong Soo Chun, Seok-jin Oh, and Sanghoon Sull School of Electrical Engineering, Korea University,
More informationTHE PROLIFERATION of multimedia material, while offering
IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 6, NO. 1, FEBRUARY 2004 103 ViBE: A Compressed Video Database Structured for Active Browsing and Search Cuneyt Taskiran, Student Member, IEEE, Jau-Yuen Chen, Alberto
More informationContent Based Image Retrieval Using Color Quantizes, EDBTC and LBP Features
Content Based Image Retrieval Using Color Quantizes, EDBTC and LBP Features 1 Kum Sharanamma, 2 Krishnapriya Sharma 1,2 SIR MVIT Abstract- To describe the image features the Local binary pattern (LBP)
More informationActive Image Database Management Jau-Yuen Chen
Active Image Database Management Jau-Yuen Chen 4.3.2000 1. Application 2. Concept This document is applied to the active image database management. The goal is to provide user with a systematic navigation
More informationResearch Article Image Retrieval using Clustering Techniques. K.S.Rangasamy College of Technology,,India. K.S.Rangasamy College of Technology, India.
Journal of Recent Research in Engineering and Technology 3(1), 2016, pp21-28 Article ID J11603 ISSN (Online): 2349 2252, ISSN (Print):2349 2260 Bonfay Publications, 2016 Research Article Image Retrieval
More informationVideo Summarization and Browsing Using Growing Cell Structures
Video Summarization and Browsing Using Growing Cell Structures Irena Koprinska and James Clark School of Information Technologies, The University of Sydney Madsen Building F09, Sydney NSW 2006, Australia
More informationA REGION-BASED APPROACH TO CONCEPTUAL IMAGE CLASSIFICATION
A REGION-BASED APPROACH TO CONCEPTUAL IMAGE CLASSIFICATION Symeon Papadopoulos 1, Vasileios Mezaris 1,2, Ioannis Kompatsiaris 2, Michael G. Strintzis 1,2 1 Information Processing Laboratory, Electrical
More informationXML Clustering by Bit Vector
XML Clustering by Bit Vector WOOSAENG KIM Department of Computer Science Kwangwoon University 26 Kwangwoon St. Nowongu, Seoul KOREA kwsrain@kw.ac.kr Abstract: - XML is increasingly important in data exchange
More informationClustering CS 550: Machine Learning
Clustering CS 550: Machine Learning This slide set mainly uses the slides given in the following links: http://www-users.cs.umn.edu/~kumar/dmbook/ch8.pdf http://www-users.cs.umn.edu/~kumar/dmbook/dmslides/chap8_basic_cluster_analysis.pdf
More information70 IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 6, NO. 1, FEBRUARY ClassView: Hierarchical Video Shot Classification, Indexing, and Accessing
70 IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 6, NO. 1, FEBRUARY 2004 ClassView: Hierarchical Video Shot Classification, Indexing, and Accessing Jianping Fan, Ahmed K. Elmagarmid, Senior Member, IEEE, Xingquan
More informationMultiView: Multilevel video content representation and retrieval*
Journal of Electronic Imaging 10(4), 895 908 (October 2001). MultiView: Multilevel video content representation and retrieval* Jianping Fan University of North Carolina at Charlotte Department of Computer
More informationAn Enhanced Image Retrieval Using K-Mean Clustering Algorithm in Integrating Text and Visual Features
An Enhanced Image Retrieval Using K-Mean Clustering Algorithm in Integrating Text and Visual Features S.Najimun Nisha 1, Mrs.K.A.Mehar Ban 2, 1 PG Student, SVCET, Puliangudi. najimunnisha@yahoo.com 2 AP/CSE,
More informationEfficient Content Based Image Retrieval System with Metadata Processing
IJIRST International Journal for Innovative Research in Science & Technology Volume 1 Issue 10 March 2015 ISSN (online): 2349-6010 Efficient Content Based Image Retrieval System with Metadata Processing
More informationVery Fast Image Retrieval
Very Fast Image Retrieval Diogo André da Silva Romão Abstract Nowadays, multimedia databases are used on several areas. They can be used at home, on entertainment systems or even in professional context
More informationClassification. Vladimir Curic. Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University
Classification Vladimir Curic Centre for Image Analysis Swedish University of Agricultural Sciences Uppsala University Outline An overview on classification Basics of classification How to choose appropriate
More informationA Novel Content Based Image Retrieval System using K-means/KNN with Feature Extraction
DOI: 10.2298/CSIS120122047C A Novel Content Based Image Retrieval System using K-means/KNN with Feature Extraction Ray-I Chang 1, Shu-Yu Lin 1,2, Jan-Ming Ho 2, Chi-Wen Fann 2, and Yu-Chun Wang 2 1 Dept.
More informationInteractive Image Retrival using Semisupervised SVM
ISSN: 2321-7782 (Online) Special Issue, December 2013 International Journal of Advance Research in Computer Science and Management Studies Research Paper Available online at: www.ijarcsms.com Interactive
More informationCluster Analysis. Ying Shen, SSE, Tongji University
Cluster Analysis Ying Shen, SSE, Tongji University Cluster analysis Cluster analysis groups data objects based only on the attributes in the data. The main objective is that The objects within a group
More informationContent Based Image Retrieval Using Hierachical and Fuzzy C-Means Clustering
Content Based Image Retrieval Using Hierachical and Fuzzy C-Means Clustering Prof.S.Govindaraju #1, Dr.G.P.Ramesh Kumar #2 #1 Assistant Professor, Department of Computer Science, S.N.R. Sons College, Bharathiar
More informationContent-Based Image Retrieval Some Basics
Content-Based Image Retrieval Some Basics Gerald Schaefer Department of Computer Science Loughborough University Loughborough, U.K. gerald.schaefer@ieee.org Abstract. Image collections are growing at a
More informationTEMPORAL AND SPATIAL SEMANTIC MODELS FOR MULTIMEDIA PRESENTATIONS ABSTRACT
TEMPORAL AND SPATIAL SEMANTIC MODELS FOR MULTIMEDIA PRESENTATIONS Shu-Ching Chen and R. L. Kashyap School of Electrical and Computer Engineering Purdue University, West Lafayette, IN 47907-285, U.S.A.
More informationAnalysis of Dendrogram Tree for Identifying and Visualizing Trends in Multi-attribute Transactional Data
Analysis of Dendrogram Tree for Identifying and Visualizing Trends in Multi-attribute Transactional Data D.Radha Rani 1, A.Vini Bharati 2, P.Lakshmi Durga Madhuri 3, M.Phaneendra Babu 4, A.Sravani 5 Department
More informationContent-based Image Retrieval using Image Partitioning with Color Histogram and Wavelet-based Color Histogram of the Image
Content-based Image Retrieval using Image Partitioning with Color Histogram and Wavelet-based Color Histogram of the Image Moheb R. Girgis Department of Computer Science Faculty of Science Minia University,
More informationImproving the Efficiency of Fast Using Semantic Similarity Algorithm
International Journal of Scientific and Research Publications, Volume 4, Issue 1, January 2014 1 Improving the Efficiency of Fast Using Semantic Similarity Algorithm D.KARTHIKA 1, S. DIVAKAR 2 Final year
More informationUnderstanding Clustering Supervising the unsupervised
Understanding Clustering Supervising the unsupervised Janu Verma IBM T.J. Watson Research Center, New York http://jverma.github.io/ jverma@us.ibm.com @januverma Clustering Grouping together similar data
More informationContent-Based Image Retrieval Using Multiple Representations
Content-Based Image Retrieval Using Multiple Representations Karin Kailing, Hans-Peter Kriegel, and Stefan Schönauer University of Munich Institute for Computer Science {kailing,kriegel,schoenauer}@dbs.ifi.lmu.de
More informationBAG-OF-VISUAL WORDS (BoVW) MODEL BASED APPROACH FOR CONTENT BASED IMAGE RETRIEVAL (CBIR) IN PEER TO PEER (P2P)NETWORKS.
BAG-OF-VISUAL WORDS (BoVW) MODEL BASED APPROACH FOR CONTENT BASED IMAGE RETRIEVAL (CBIR) IN PEER TO PEER (P2P)NETWORKS. 1 R.Lavanya, 2 E.Lavanya, 1 PG Scholar, Dept Of Computer Science Engineering,Mailam
More informationA Content Based Image Retrieval System Based on Color Features
A Content Based Image Retrieval System Based on Features Irena Valova, University of Rousse Angel Kanchev, Department of Computer Systems and Technologies, Rousse, Bulgaria, Irena@ecs.ru.acad.bg Boris
More informationContent Based Image Retrieval with Semantic Features using Object Ontology
Content Based Image Retrieval with Semantic Features using Object Ontology Anuja Khodaskar Research Scholar College of Engineering & Technology, Amravati, India Dr. S.A. Ladke Principal Sipna s College
More informationDifferential Compression and Optimal Caching Methods for Content-Based Image Search Systems
Differential Compression and Optimal Caching Methods for Content-Based Image Search Systems Di Zhong a, Shih-Fu Chang a, John R. Smith b a Department of Electrical Engineering, Columbia University, NY,
More information5/15/16. Computational Methods for Data Analysis. Massimo Poesio UNSUPERVISED LEARNING. Clustering. Unsupervised learning introduction
Computational Methods for Data Analysis Massimo Poesio UNSUPERVISED LEARNING Clustering Unsupervised learning introduction 1 Supervised learning Training set: Unsupervised learning Training set: 2 Clustering
More informationData Clustering Hierarchical Clustering, Density based clustering Grid based clustering
Data Clustering Hierarchical Clustering, Density based clustering Grid based clustering Team 2 Prof. Anita Wasilewska CSE 634 Data Mining All Sources Used for the Presentation Olson CF. Parallel algorithms
More informationClustering Methods for Video Browsing and Annotation
Clustering Methods for Video Browsing and Annotation Di Zhong, HongJiang Zhang 2 and Shih-Fu Chang* Institute of System Science, National University of Singapore Kent Ridge, Singapore 05 *Center for Telecommunication
More informationK-Means Clustering Using Localized Histogram Analysis
K-Means Clustering Using Localized Histogram Analysis Michael Bryson University of South Carolina, Department of Computer Science Columbia, SC brysonm@cse.sc.edu Abstract. The first step required for many
More informationIMAGE RETRIEVAL: A STATE OF THE ART APPROACH FOR CBIR
IMAGE RETRIEVAL: A STATE OF THE ART APPROACH FOR CBIR AMANDEEP KHOKHER Department of ECE, RIMT-MAEC Mandi Gobindgarh, Punjab, India amandeep.khokher@gmail.com DR. RAJNEESH TALWAR Department of ECE, RIMT-IET
More informationSemantic Extraction and Semantics-based Annotation and Retrieval for Video Databases
Semantic Extraction and Semantics-based Annotation and Retrieval for Video Databases Yan Liu (liuyan@cs.columbia.edu) and Fei Li (fl200@cs.columbia.edu) Department of Computer Science, Columbia University
More informationImage Matching Using Run-Length Feature
Image Matching Using Run-Length Feature Yung-Kuan Chan and Chin-Chen Chang Department of Computer Science and Information Engineering National Chung Cheng University, Chiayi, Taiwan, 621, R.O.C. E-mail:{chan,
More informationImage Retrieval by Example: Techniques and Demonstrations
Image Retrieval by Example: Techniques and Demonstrations Elena Stringa, Paul Meylemans 1 and João G.M. Gonçalves Joint Research Centre, 21020 Ispra (VA), ITALY 1 Euratom Safeguards Office, LUXEMBOURG
More informationA B2B Search Engine. Abstract. Motivation. Challenges. Technical Report
Technical Report A B2B Search Engine Abstract In this report, we describe a business-to-business search engine that allows searching for potential customers with highly-specific queries. Currently over
More informationSimilarity of Medical Images Computed from Global Feature Vectors for Content-based Retrieval
Similarity of Medical Images Computed from Global Feature Vectors for Content-based Retrieval Thomas M. Lehmann 1, Mark O. Güld 1, Daniel Keysers 2, Thomas Deselaers 2, Henning Schubert 3, Berthold Wein
More informationFIRE Flexible Image Retrieval Engine: ImageCLEF 2004 Evaluation
FIRE Flexible Image Retrieval Engine: ImageCLEF 2004 Evaluation Thomas Deselaers, Daniel Keysers, and Hermann Ney Lehrstuhl für Informatik VI Computer Science Department, RWTH Aachen University D-52056
More information10701 Machine Learning. Clustering
171 Machine Learning Clustering What is Clustering? Organizing data into clusters such that there is high intra-cluster similarity low inter-cluster similarity Informally, finding natural groupings among
More informationImage Retrieval System Based on Sketch
Image Retrieval System Based on Sketch Author 1 Mrs. Asmita A. Desai Assistant Professor,Department of Electronics Engineering, Author 2 Prof. Dr. A. N. Jadhav HOD,Department of Electronics Engineering,
More informationSpeeding up Queries in a Leaf Image Database
1 Speeding up Queries in a Leaf Image Database Daozheng Chen May 10, 2007 Abstract We have an Electronic Field Guide which contains an image database with thousands of leaf images. We have a system which
More informationAn Analysis of Image Retrieval Behavior for Metadata Type and Google Image Database
An Analysis of Image Retrieval Behavior for Metadata Type and Google Image Database Toru Fukumoto Canon Inc., JAPAN fukumoto.toru@canon.co.jp Abstract: A large number of digital images are stored on the
More informationStill Image Objective Segmentation Evaluation using Ground Truth
5th COST 276 Workshop (2003), pp. 9 14 B. Kovář, J. Přikryl, and M. Vlček (Editors) Still Image Objective Segmentation Evaluation using Ground Truth V. Mezaris, 1,2 I. Kompatsiaris 2 andm.g.strintzis 1,2
More informationA Miniature-Based Image Retrieval System
A Miniature-Based Image Retrieval System Md. Saiful Islam 1 and Md. Haider Ali 2 Institute of Information Technology 1, Dept. of Computer Science and Engineering 2, University of Dhaka 1, 2, Dhaka-1000,
More informationAn Image Data Model. William I. Grosky and Peter L. Stanchev
An Image Data Model William I. Grosky and Peter L. Stanchev Department of Computer Science, Wayne State University Detroit, MI 48202 {grosky, stanchev}@cs.wayne.edu Abstract. In this paper, we analyze
More informationInteroperable Content-based Access of Multimedia in Digital Libraries
Interoperable Content-based Access of Multimedia in Digital Libraries John R. Smith IBM T. J. Watson Research Center 30 Saw Mill River Road Hawthorne, NY 10532 USA ABSTRACT Recent academic and commercial
More informationA Novel Image Retrieval Method Using Segmentation and Color Moments
A Novel Image Retrieval Method Using Segmentation and Color Moments T.V. Saikrishna 1, Dr.A.Yesubabu 2, Dr.A.Anandarao 3, T.Sudha Rani 4 1 Assoc. Professor, Computer Science Department, QIS College of
More informationWelcome Back to Fundamental of Multimedia (MR412) Fall, ZHU Yongxin, Winson
Welcome Back to Fundamental of Multimedia (MR412) Fall, 2012 ZHU Yongxin, Winson zhuyongxin@sjtu.edu.cn Content-Based Retrieval in Digital Libraries 18.1 How Should We Retrieve Images? 18.2 C-BIRD : A
More informationContent Based Image Retrieval (CBIR) Using Segmentation Process
Content Based Image Retrieval (CBIR) Using Segmentation Process R.Gnanaraja 1, B. Jagadishkumar 2, S.T. Premkumar 3, B. Sunil kumar 4 1, 2, 3, 4 PG Scholar, Department of Computer Science and Engineering,
More informationAN IMPROVED K-MEANS CLUSTERING ALGORITHM FOR IMAGE SEGMENTATION
AN IMPROVED K-MEANS CLUSTERING ALGORITHM FOR IMAGE SEGMENTATION WILLIAM ROBSON SCHWARTZ University of Maryland, Department of Computer Science College Park, MD, USA, 20742-327, schwartz@cs.umd.edu RICARDO
More informationA Novel Criterion Function in Feature Evaluation. Application to the Classification of Corks.
A Novel Criterion Function in Feature Evaluation. Application to the Classification of Corks. X. Lladó, J. Martí, J. Freixenet, Ll. Pacheco Computer Vision and Robotics Group Institute of Informatics and
More informationColor interest points detector for visual information retrieval
Color interest points detector for visual information retrieval Jérôme Da Rugna a and Hubert Konik a a LIGIV, Université Jean Monnet, 3 rue Javelin Pagnon, 42000, St Etienne, France ABSTRACT The purpose
More informationCluster Analysis. Mu-Chun Su. Department of Computer Science and Information Engineering National Central University 2003/3/11 1
Cluster Analysis Mu-Chun Su Department of Computer Science and Information Engineering National Central University 2003/3/11 1 Introduction Cluster analysis is the formal study of algorithms and methods
More informationQuerying by Color Regions using the VisualSEEk Content-Based Visual Query System
Querying by Color Regions using the VisualSEEk Content-Based Visual Query System John R. Smith and Shih-Fu Chang Center for Image Technology for New Media and Department of Electrical Engineering Columbia
More informationStats 170A: Project in Data Science Exploratory Data Analysis: Clustering Algorithms
Stats 170A: Project in Data Science Exploratory Data Analysis: Clustering Algorithms Padhraic Smyth Department of Computer Science Bren School of Information and Computer Sciences University of California,
More informationSegmentation Computer Vision Spring 2018, Lecture 27
Segmentation http://www.cs.cmu.edu/~16385/ 16-385 Computer Vision Spring 218, Lecture 27 Course announcements Homework 7 is due on Sunday 6 th. - Any questions about homework 7? - How many of you have
More informationCS 1675 Introduction to Machine Learning Lecture 18. Clustering. Clustering. Groups together similar instances in the data sample
CS 1675 Introduction to Machine Learning Lecture 18 Clustering Milos Hauskrecht milos@cs.pitt.edu 539 Sennott Square Clustering Groups together similar instances in the data sample Basic clustering problem:
More informationText Enhancement with Asymmetric Filter for Video OCR. Datong Chen, Kim Shearer and Hervé Bourlard
Text Enhancement with Asymmetric Filter for Video OCR Datong Chen, Kim Shearer and Hervé Bourlard Dalle Molle Institute for Perceptual Artificial Intelligence Rue du Simplon 4 1920 Martigny, Switzerland
More informationBy choosing to view this document, you agree to all provisions of the copyright laws protecting it.
Jussi Pakkanen and Jukka Iivarinen, A Novel Self Organizing Neural Network for Defect Image Classification. In Proceedings of the International Joint Conference on Neural Networks, pages 2553 2558, Budapest,
More informationFish species recognition from video using SVM classifier
Fish species recognition from video using SVM classifier Katy Blanc, Diane Lingrand, Frédéric Precioso Univ. Nice Sophia Antipolis, I3S, UMR 7271, 06900 Sophia Antipolis, France CNRS, I3S, UMR 7271, 06900
More informationExtended Query Refinement for Content-Based Access to Large Medical Image Databases
Extended Query Refinement for Content-Based Access to Large Medical Image Databases Thomas M. Lehmann 1, Bartosz Plodowski, Klaus Spitzer, Berthold B. Wein a, Hermann Ney b, Thomas Seidl c Department of
More informationColour Segmentation-based Computation of Dense Optical Flow with Application to Video Object Segmentation
ÖGAI Journal 24/1 11 Colour Segmentation-based Computation of Dense Optical Flow with Application to Video Object Segmentation Michael Bleyer, Margrit Gelautz, Christoph Rhemann Vienna University of Technology
More informationContent based Image Retrieval Using Multichannel Feature Extraction Techniques
ISSN 2395-1621 Content based Image Retrieval Using Multichannel Feature Extraction Techniques #1 Pooja P. Patil1, #2 Prof. B.H. Thombare 1 patilpoojapandit@gmail.com #1 M.E. Student, Computer Engineering
More informationAUTOMATIC IMAGE ANNOTATION AND RETRIEVAL USING THE JOINT COMPOSITE DESCRIPTOR.
AUTOMATIC IMAGE ANNOTATION AND RETRIEVAL USING THE JOINT COMPOSITE DESCRIPTOR. Konstantinos Zagoris, Savvas A. Chatzichristofis, Nikos Papamarkos and Yiannis S. Boutalis Department of Electrical & Computer
More informationJPEG 2000 Based Scalable Summary for Understanding Long Video Surveillance Sequences
JPEG 2000 Based Scalable Summary for Understanding Long Video Surveillance Sequences Jerome Meessen a, Jean-Francois Delaigle a, Li-Qun Xu b and Benoit Macq c. a Multitel asbl, 7000 Mons, Belgium {jerome.meessen,
More informationInternational Journal of Computer Engineering and Applications, Volume XII, Special Issue, March 18, ISSN
USABILITY OF MAPREDUCE NEURAL NETWORK FRAMEWORK IN IMAGE RECOGNITION 1 Dr. Sitalakshmi Venkatraman 2 Dr. Siddhivinayak Kulkarni 1 Department of Information Technology, Melbourne Polytechnic, Prahran, Australia
More informationClustering. CS294 Practical Machine Learning Junming Yin 10/09/06
Clustering CS294 Practical Machine Learning Junming Yin 10/09/06 Outline Introduction Unsupervised learning What is clustering? Application Dissimilarity (similarity) of objects Clustering algorithm K-means,
More information2 Line-Angle-Ratio Statistics Experiments on various types of images showed us that one of the strongest spatial features of an image is its line segm
Using Texture in Image Similarity and Retrieval Selim Aksoy and Robert M. Haralick Intelligent Systems Laboratory Department of Electrical Engineering University of Washington Seattle, WA 9895-25 faksoy,haralickg@isl.ee.washington.edu
More informationDominant colour extraction in DCT domain
Image and Vision Computing 24 (2006) 1269 1277 www.elsevier.com/locate/imavis Dominant colour extraction in DCT domain Jianmin Jiang a,b, *, Ying Weng b, PengJie Li c a Southwest University, China b University
More informationContent-Based Image Retrieval with LIRe and SURF on a Smartphone-Based Product Image Database
Content-Based Image Retrieval with LIRe and SURF on a Smartphone-Based Product Image Database Kai Chen 1 and Jean Hennebert 2 1 University of Fribourg, DIVA-DIUF, Bd. de Pérolles 90, 1700 Fribourg, Switzerland
More informationContent Based Image Retrieval Using Combined Color & Texture Features
IOSR Journal of Electrical and Electronics Engineering (IOSR-JEEE) e-issn: 2278-1676,p-ISSN: 2320-3331, Volume 11, Issue 6 Ver. III (Nov. Dec. 2016), PP 01-05 www.iosrjournals.org Content Based Image Retrieval
More informationObject-Based Image Labeling through Learning-by-Example and Multi-Level Segmentation. Rochester, NY 14627, USA
Object-Based Image Labeling through Learning-by-Example and Multi-Level Segmentation Y. Xu (), P. Duygulu (2), E. Saber (), (3), A. M. Tekalp (), (4), and F. T. Yarman-Vural (2) (). Department of Electrical
More informationIterative Image Based Video Summarization by Node Segmentation
Iterative Image Based Video Summarization by Node Segmentation Nalini Vasudevan Arjun Jain Himanshu Agrawal Abstract In this paper, we propose a simple video summarization system based on removal of similar
More informationAS STORAGE and bandwidth capacities increase, digital
974 IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 13, NO. 7, JULY 2004 Concept-Oriented Indexing of Video Databases: Toward Semantic Sensitive Retrieval and Browsing Jianping Fan, Hangzai Luo, and Ahmed
More informationAutomatic Video Caption Detection and Extraction in the DCT Compressed Domain
Automatic Video Caption Detection and Extraction in the DCT Compressed Domain Chin-Fu Tsao 1, Yu-Hao Chen 1, Jin-Hau Kuo 1, Chia-wei Lin 1, and Ja-Ling Wu 1,2 1 Communication and Multimedia Laboratory,
More informationHierarchical Clustering Lecture 9
Hierarchical Clustering Lecture 9 Marina Santini Acknowledgements Slides borrowed and adapted from: Data Mining by I. H. Witten, E. Frank and M. A. Hall 1 Lecture 9: Required Reading Witten et al. (2011:
More informationShort Run length Descriptor for Image Retrieval
CHAPTER -6 Short Run length Descriptor for Image Retrieval 6.1 Introduction In the recent years, growth of multimedia information from various sources has increased many folds. This has created the demand
More informationA Perceptual Model Based on Computational Features for Texture Representation and Retrieval
A Perceptual Model Based on Computational Features for Texture Representation and Retrieval 1 K.N.Sindhuri, 2 N.Leelavathy, 3 B.Srinivas 1,2,3 Dept. of CSE, Pragati Engineering College, Surampalem, AP,
More informationEXPLORING ON STEGANOGRAPHY FOR LOW BIT RATE WAVELET BASED CODER IN IMAGE RETRIEVAL SYSTEM
TENCON 2000 explore2 Page:1/6 11/08/00 EXPLORING ON STEGANOGRAPHY FOR LOW BIT RATE WAVELET BASED CODER IN IMAGE RETRIEVAL SYSTEM S. Areepongsa, N. Kaewkamnerd, Y. F. Syed, and K. R. Rao The University
More informationColor Image Segmentation
Color Image Segmentation Yining Deng, B. S. Manjunath and Hyundoo Shin* Department of Electrical and Computer Engineering University of California, Santa Barbara, CA 93106-9560 *Samsung Electronics Inc.
More informationMultimodal Information Spaces for Content-based Image Retrieval
Research Proposal Multimodal Information Spaces for Content-based Image Retrieval Abstract Currently, image retrieval by content is a research problem of great interest in academia and the industry, due
More informationBuilding a Concept Hierarchy from a Distance Matrix
Building a Concept Hierarchy from a Distance Matrix Huang-Cheng Kuo 1 and Jen-Peng Huang 2 1 Department of Computer Science and Information Engineering National Chiayi University, Taiwan 600 hckuo@mail.ncyu.edu.tw
More information